How to Use Muse to Investigate Options

Yes — open-ended 'what are my options?' questions are where an agent beats a search box. Muse, Meta's personal AI agent, can browse with its own browser to map the landscape: what the paths are, what each costs in money and time, which prerequisites apply, and what people who chose each path say afterward. It works asynchronously, so a broad sweep can finish overnight, and it remembers your situation, so the shortlist reflects your constraints rather than a generic top-ten list. Ask for a decision memo with sources and you get something you can act on — after you verify the load-bearing facts yourself.

• The broader the question, the more room for error. Press reporting from launch week documents Muse sometimes failing to finish tasks — scope the investigation and check its sources.
• Reviews and forums are inputs, not verdicts. Muse can quote them, but selection bias runs deep in every 'was it worth it?' thread on the internet.

1. 1

   Describe your situation before your question

   'Ten years in retail management, no degree, evenings only' changes which options exist at all. Muse remembers it for the whole investigation.

2. 2

   Ask for the map before the recommendation

   First run: enumerate every option category. Only then go deep on each. Recommendations built on unmapped landscapes miss things.

3. 3

   Set the vetting criteria

   Cost, time, prerequisites, realistic outcomes, reversibility. Tell Muse what matters and have it score every option against the same yardstick.

4. 4

   Require receipts

   Ask for a source next to every factual claim — program costs, requirement lists, outcome figures — so verification takes minutes, not evenings.

A retail store manager without a degree is investigating whether an online data-analytics certificate could move them into an analyst role within a year.

Goal: Map and vet my options for moving from retail management into data work within 12 months.
Context: I manage a store, 50 hours a week, no degree, $3,000 budget, evenings and Sundays only. I'm leaning toward an online analytics certificate, but pressure-test that.
Requirements:
- Round one: every distinct path — certificates, community college, self-teaching, an internal transfer at my company, adjacent roles.
- Round two for the shortlist: total cost, hours per week, prerequisites, real hiring outcomes, reversibility.
- Same criteria for every option.
Actions:
- Check official program pages and labor statistics first, then forums for people who actually finished these programs.
- Attach a source to every claim; mark gaps as NEEDS CHECK.
Approval: Research only — no inquiry forms or contact with program salespeople without asking me.
Output: Option map, vetted shortlist with sources, an honest paragraph on my certificate lean, and the NEEDS CHECK list.

• Your full situation

  Time, money, family, current role — options that ignore constraints are daydreams, and Muse's memory keeps the constraints applied throughout.

• Non-negotiables

  These kill options quickly and cheaply, before anyone falls in love with a brochure.

• Your current leaning

  Telling Muse your bias lets it pressure-test the favorite instead of mirroring it.

• A decision date

  Investigations expand to fill available time. A deadline forces the two-round structure that actually ends in an answer.

• What success would look like

  'A better job' and '$15k more within two years' produce completely different shortlists.

For this task, it is smart to make Muse check with you before it takes certain actions. Decide your rules up front and include them in the prompt.

• Submitting inquiry or request-information forms — These mint sales calls. Muse checks before sensitive actions like submitting forms — keep that gate shut until you are ready.

• Paying application or deposit fees — Application fees are real money flowing into a pipeline that is awkward to reverse. Approve each one individually.

• Talking to admissions or sales reps — Their job is conversion. The researched brief should arm you first, not follow the pitch.

• Acting on outcome claims without checking sources — Salary and placement numbers are marketing-adjacent. Verify the load-bearing ones at the cited source yourself.

• ✓Insist on the option map before any ranking. The best path is sometimes a category you never named, and an agent with a browser will find it if you ask.
• ✓Make reversibility an explicit criterion. It reframes close calls: the option you can undo cheaply beats the marginally better one you cannot.
• ✓Ask Muse to argue against your leaning in one paragraph. Its memory of your situation makes the counterargument specific instead of generic.
• ✓Two rounds beat one: a broad, cheap sweep first, then deep vetting only on the survivors. Long single runs are where new agents trail off.
• ✓Set the source order explicitly: official pages for facts, forums for texture. Left to its own devices, Muse may weight them equally.